HIGH THROUGHPUT SNP IDENTIFICATION. LNG has discovered over 100 single nucleotide polymorphisms (SNPs) in more than 50 neuro-genetic candidate genes for addictive behaviors, including alcoholism. Approximately 20% of the SNPs discovered by LNG result in non-synonymous amino acid changes that may alter function of the encoded protein or alterations in expression of the gene. The implications of detected sequence variation have been profound: for example a rare serotonin transporter variant we found was shown to be functional and discovered to lead to severe pathology: Asperger's syndrome, treatment resistant OCD, and anorexia nervosa, in two families in which it is segregating. For example, the common HTR2C Ser23Cys and HTR2A Asn452His alleles have been shown to be functional and linked to the clozapine responsiveness of schizophrenics. Genes for SNP screening were selected on the basis of postulated roles in alcohol, treatment response, identification in whole genome or candidate linkage or association study, and the availability of genomic sequence data. Because a goal of the project is to determine the role of SNPs in complex genetic disorders, we focused our screening in efforts on protein coding portions of the candidate gene and possible regulatory regions within and flanking the protein coding regions. We used denaturing high performance liquid chromatography (dHPLC), which detects greater than 90% of sequence variants in DNA duplexes of 150-450 bp in size, to screen a DNA panel composed of 477 genomic DNAs enriched for clinical and ethnic diversity. Size and diversity of the screening panel are critical. For example, a rare serotonin transport missense substitution was found in only two families, however individuals with the substitution have severe behavioral problems. Except in one case, they have Asperger's Syndrome, severe treatment resistant OCD, or anorexia nervosa. The screening panel was constructed in order to achieve an accurate estimation of allele frequencies for each novel SNP within our primary linkage data sets. The screening dataset is composed of native American, African American, Caucasian and Asian populations and is enriched for various psychiatric diseases including alcoholism, OCD, schizophrenia, eating disorders, and bipolar disorder. Sequence variants obtained from the primary screening method were characterized and confirmed using gel-based, semi-automated, or capillary-based, fully automated, DNA sequencing methods. The continuing requirement for SNP detection using accurate, high throughput, and lower cost methods motivated us to develop other methods. Our experimental approach took advantage of thermal melting differences between double stranded DNA molecules that carry either a complete sequence match or a single nucleotide mismatch, as found in heterozygous individuals for a particular SNP. Denaturation and re-annealing of the strands in a DNA mixture from an individual that is heterozygous for a SNP will contain at most 50% of the re-annealed DNAs as perfectly matched sequences, while at least 50% of the re-annealed DNAs will have a single base mismatch. The thermal melting characteristics of DNA duplexes with a single base mismatch are inherently less stable than perfectly matched double stranded DNAs. A Perkin Elmer 7700 Sequence Detector was used to monitor the change in fluorescence signal over the course of double strand DNA melting. With this particular instrument, it was possible to perform thermal denaturation experiments & analyze the DNA melting data in 96 samples at one time. We transformed the raw fluorescence readings into first derivative plots that produced a melting profile for each DNA duplex. With these melting profiles, we were able to detect single-base mismatches in double stranded DNAs 100-150 bp in size. This approach is highly suited for large-scale detection of new sequence variants. HIGH THROUGHPUT GENOTYPING. Central issues in high throughput genotyping procedures are accuracy, flexibility, and cost. Because of its assay design flexibility, low error rate, and potential for performing 96 to 384 assays simultaneously, we selected the 5' nuclease assay using SNPs either discovered in the LNG or SNPs that have been described by other laboratories for allele frequency determinations in our study populations. Genotyping completion rates are >95% and error rates that have been calculated from replicate samples are less than 0.5%. Reagent costs are currently $0.20-$0.30 per assay. We are now transitioning to a higher-throughput, multiplex genotyping method utilizing DNA ligation, with detection on an ABI capillary sequencer.